No abstract

A new strategy for detection and diagnosis of broken bars and load oscillations is proposed. The proposed technique is based on the analysis of the instantaneous active and reactive powers spectra, and allows the correct distinction between broken rotor bars and low-frequency load oscillations. A suitable fault severity factor is also proposed, which allows the broken rotor bar fault quantification, while being practically independent of the motor load. By using the proposed technique, a broken bar fault can be detected and diagnosed, even in the presence of load oscillations at frequencies next to twice the slip frequency. Extensive experimental results, both from laboratory and industrial cases, validate the proposal.

No abstract

No abstract

In this paper, a model-based strategy for statorinterturn short-circuit detection on induction motors is presented. The proposed strategy is based on the generation of a vector of specific residual using a state observer. The vectorial residual is generated from a decomposition of the current estimation error. This allows for a fast detection of incipient faults, independently of the phase in which the fault occurs. Since the observer includes an adaptive scheme for rotor-speed estimation, the proposed scheme can be implemented for online monitoring, by measuring only stator voltages and currents. It is shown that the proposed strategy presents very low sensitivity to load variations and power-supply perturbations. Experimental results are included to show the ability of the proposed strategy for detecting incipient faults, including a low number of short-circuited turns and low fault current.

A new method to calculate the inductances of induction machines considering axial nonuniformity is proposed. The proposed method, an extension of the modified winding function approach, allows considering nonuniformity due to skew and static and dynamic air-gap eccentricity. Theoretical fundamentals and experimental results that validate the proposed method are presented.

No abstract

No abstract